Deep learning and big data mining for Metal-Organic frameworks with high performance for simultaneous desulfurization and carbon capture

Kexin Guan; Fangyi Xu; Xiaoshan Huang; Yu Li; Shuya Guo; Yizhen Situ; You Chen; Jianming Hu; Zili Liu; Hong Liang; Xin Zhu; Yufang Wu; Zhiwei Qiao

doi:10.1016/j.jcis.2024.02.098

Deep learning and big data mining for Metal-Organic frameworks with high performance for simultaneous desulfurization and carbon capture

J Colloid Interface Sci. 2024 May 15:662:941-952. doi: 10.1016/j.jcis.2024.02.098. Epub 2024 Feb 15.

Authors

Kexin Guan¹, Fangyi Xu¹, Xiaoshan Huang¹, Yu Li¹, Shuya Guo¹, Yizhen Situ², You Chen¹, Jianming Hu³, Zili Liu¹, Hong Liang¹, Xin Zhu⁴, Yufang Wu⁵, Zhiwei Qiao⁶

Affiliations

¹ Guangzhou Key Laboratory for New Energy and Green Catalysis, School of Chemistry and Chemical Engineering, Guangzhou University, Guangzhou 510006, China.
² Guangzhou Key Laboratory for New Energy and Green Catalysis, School of Chemistry and Chemical Engineering, Guangzhou University, Guangzhou 510006, China; State Key Laboratory of Organic-Inorganic Composites, Beijing University of Chemical Technology, Beijing 100029, China.
³ College of Economics and Statistics, Guangzhou University, Guangzhou 510006, China.
⁴ Guangzhou Key Laboratory for New Energy and Green Catalysis, School of Chemistry and Chemical Engineering, Guangzhou University, Guangzhou 510006, China; College of Economics and Statistics, Guangzhou University, Guangzhou 510006, China. Electronic address: zhux@gzhu.edu.cn.
⁵ Guangzhou Key Laboratory for New Energy and Green Catalysis, School of Chemistry and Chemical Engineering, Guangzhou University, Guangzhou 510006, China. Electronic address: yufang.wu@gzhu.edu.cn.
⁶ Guangzhou Key Laboratory for New Energy and Green Catalysis, School of Chemistry and Chemical Engineering, Guangzhou University, Guangzhou 510006, China. Electronic address: zqiao@gzhu.edu.cn.

PMID: 38382377
DOI: 10.1016/j.jcis.2024.02.098

Abstract

Carbon capture and desulfurization of flue gases are crucial for the achievement of carbon neutrality and sustainable development. In this work, the "one-step" adsorption technology with high-performance metal-organic frameworks (MOFs) was proposed to simultaneously capture the SO₂ and CO₂. Four machine learning algorithms were used to predict the performance indicators (N_CO2+SO2, S_CO2+SO2/N2, and TSN) of MOFs, with Multi-Layer Perceptron Regression (MLPR) showing better performance (R² = 0.93). To address sparse data of MOF chemical descriptors, we introduced the Deep Factorization Machines (DeepFM) model, outperforming MLPR with a higher R² of 0.95. Then, sensitivity analysis was employed to find that the adsorption heat and porosity were the key factors for SO₂ and CO₂ capture performance of MOF, while the influence of open alkali metal sites also stood out. Furthermore, we established a kinetic model to batch simulate the breakthrough curves of TOP 1000 MOFs to investigate their dynamic adsorption separation performance for SO₂/CO₂/N₂. The TOP 20 MOFs screened by the dynamic performance highly overlap with those screened by the static performance, with 76 % containing open alkali metal sites. This integrated approach of computational screening, machine learning, and dynamic analysis significantly advances the development of efficient MOF adsorbents for flue gas treatment.

Keywords: Adsorption separation; Breakthrough curves; Deep learning; Metal–organic frameworks; Open metal sites.